This week I set out for myself as a sort of oh-my-God-what-am-I-doing week, and also to make me a couple of style tests. I had vague ideas floating around in my head about what to do, and getting them out of my system was definitely the way to go. I feel like I’m a little more ready on my feet now. Here are the two style tests;
The simplistic style test was very quick and easy to make, taking half the time of the PBR piece. I really like where it’s going, however I feel that it looks almost too simple, and I don’t think the style would hold up in a larger diorama. The PBR diorama has more to it both visually and technically, and I feel as if it would lend itself better to a lively, animated environment like I intend on creating. Also, the more complex style shows I’m capable of more than just hand-painting albedo textures.
I feel that a simplistic style would be good to use in a personal project alongside my FMP to keep things diverse and have more unusual projects in my portfolio such as the forest gif I created during summer, which has attracted a lot of attention.
There are a few things that I will need to work on to make the PBR style more suitable for my FMP dioramas. Firstly, it’s extremely noisy and in some places confusing to look at. I’ll need to use composition, lighting, colour, and texture complexity to control where the viewer looks. I’ve achieved this better in the simplistic style test. Second, I managed to not use Zbrush for the entire PBR diorama by creating height maps in Photoshop. For some assets, this worked fine but for others they look really strange up close. I’ll need to use Zbrush for some assets. Leading on from that, the worst asset of all is the rock, which would also benefit from Zbrush for the overall lowpoly shape as well as the normal mapped details. This will stop it looking like I’ve taken a blob and applied a flat rock texture to it.
Overall I’ve had a very mixed reception for my two style tests. It’s about 50/50 for which one people prefer. I took my work to Polycount, and 3 out of the 4 people who have replied so far prefer the left-hand one. But everyone else who has given their opinion prefers the right. The main arguments for the simplistic test is that it’s more unusual and would stand out stylistically, and that it’s easier on the eye. I prefer the PBR test because it is more of a challenge for me to create, has more going on, and I think I’ll be prouder of myself at the end. One stand out point against the PBR style is that it will take me longer to create assets, and so I will achieve less. With this and the other considerations in mind, I think it will be sensible to create a marriage of the two.
Though it’s felt weird doing work that doesn’t physically contribute to my FMP hand-in work, I know that I’ve saved myself a lot of long-term frustration. Since finishing my tests, I’ve made a list of things I’ve learned thanks to the work I’ve done this week.
- Flowers need to be more dense to withstand camera distance.
- Grass blades will need a quick normal map.
- Certain assets such as trees, rocks, and table will need Zbrushing.
- I shouldn’t be stingy on tris.
- Avoid creating too much noise with leaves, lighting, textures etc.
- Need to more carefully control lighting to create a focus.
- Will need dirt ground texture as well as grass.
- Need to be careful to create ground textures that look ok with grass blades sprouting from them.
I’ve also updated a list of things I need to learn in UE4 and made tracks on finding documentation to help me with those things. I’ve been testing some shaders and animations to speed up the process down the line, meaning I’ll be able to spend more time on the art than fiddling with technical stuff.
Finally, I’ve been playing The Witcher 3 and examining how the plant life was created; I took over 70 screenshots. I learned some really useful tricks that I’ll be testing on trees and flowers that I make. Most striking for me was the fact that the trees in The Witcher, made by Speedtree, don’t just have wavy leaves, but the trunks bow in the wind too. It looks awesome and adds life and diversity to the world. Also, all the trees have a mixture of fixed leaf planes, and ones that turn to face the player as they move so that the tree canopy looks dense from all angles. This is something I wish to try.
I’ve now created an Excel spreadsheet I can fill out on a per-diorama basis, so next week I can create my diorama concept, write out every asset, animation etc., and then allocate my time over the course of the 5 weeks. This method of planning my time is something I developed over the course of my Style Matrix projects, and I feel works best for me.
So, by the end of next week I intend to have a final concept, and to have started a blockout for the final diorama. From there I can begin creating the finalised assets, animations, particles etc. for the scene.